Cooperation of germ line JAK2 mutations E846D and R1063H in hereditary erythrocytosis with megakaryocytic atypia

Somatic JAK2 mutations are the most common disease-causing event in patients with myeloproliferative neoplasms (MPNs). In the majority of MPN patients, an acquired gain-of-function V617F JAK2 mutation leads to constitutive activation of the JAK2 kinase and subsequently to excessive activation of JAK/STAT signaling.^1-4  Recently, germ line JAK2 mutations (different from V617F substitution) have been described in cases with familial MPNs exhibiting hereditary polyclonal thrombocytosis^5-7  and triple-negative MPNs.⁸  All of these inherited JAK2 mutations, localized in the kinase and pseudokinase domain of JAK2, signal through the thrombopoietin receptor (MPL) rather than the erythropoietin (EPO) receptor (EPOR), explaining the phenotype of the polyclonal disease. It is proposed that differential signaling of STATs leads to different clinical phenotypes associated with different activating JAK2 mutations: essential thrombocythemia and hereditary thrombocytosis being promoted by MPL/STAT1 signaling; STAT5 activation resulting in excessive erythroid proliferation and polycythemia vera (PV).^6,9 

Here, we studied a patient who inherited 2 JAK2 mutations and presented with erythrocytosis and abnormal megakaryopoiesis in the bone marrow (BM), partially resembling PV cases with JAK2 exon 12 mutations.¹⁰  Thus, in contrast to previously studied germ line JAK2 mutations, which were all heterozygous, here mutations on both JAK2 alleles are present. We show that both of these mutations are activating in the context of EPOR, but with different characteristics. Functional analyses of the 2 germ line JAK2 mutations suggested the threshold of JAK2 activity necessary for induction of erythrocytosis.

Descriptions of assays and tests performed on patient’s samples can be found in supplemental Materials and methods (available on the Blood Web site).

Ba/F3-EPOR¹¹  JAK2 stable transfectants or γ-2A cells were used in viability and proliferation assays, analyses of JAK2 signal transduction, and dual-luciferase assays for STAT transcriptional activities. Retrovirally transduced mouse BM cells were used for in vitro colony assay. Analysis of x-ray crystallography data was performed using PyMOL software. For detailed descriptions of individual procedures, see supplemental Materials and methods.

The diagnosis of polycythemia in a 15-year-old boy was based on the hematological and clinical data summarized in Figure 1A. The patient had increased hemoglobin/hematocrit level, higher red blood cell count, plethora, palpable spleen, and normal EPO level. Although the white blood cell differential and platelet count were normal (supplemental Table 1), the BM aspirate and BM biopsy showed hypercellularity, erythroid hyperplasia, and atypical, enlarged, and immature megakaryocytes, likely resulting from the stimulation of the megakaryocytic lineage (Figure 1B; supplemental Figure 1). The iron status parameters were normal (supplemental Table 2). The in vitro colony assay revealed hypersensitivity of the patient’s burst-forming unit-erythroid (BFU-E) progenitors to EPO (Figure 1C), a feature characteristic for primary polycythemia states.¹²  Both parents and the patient’s brother were clinically normal, with normal EPO levels and hematological (Figure 1A) and iron status parameters, with the exception of elevated ferritin in the father and brother likely due to infection (supplemental Table 2). Nevertheless, the sensitivity of BFU-E progenitors to EPO was significantly increased in the mother and slightly also in the father compared with healthy controls and the patient’s brother (Figure 1C). In addition, the colony assay revealed significantly increased numbers of circulating BFU-E progenitors in the patient and his parents compared with healthy controls (supplemental Figure 2A). These results suggested the hereditary nature of the propositus’s erythrocytosis. No somatic mutations or chromosomal aberrations were uncovered by either whole-exome sequencing or microarray analysis (supplemental Table 3i-ii), suggesting that the hematopoiesis was likely polyclonal. Targeted mutational screening (see supplemental Materials and methods) and search for germ line variants using whole-exome sequencing revealed a E846D substitution in JAK2 in a heterozygous state in the propositus and his mother and JAK2 R1063H in the propositus and his father (supplemental Results; supplemental Figure 2B); both variants were reported as polymorphisms with very low frequencies (supplemental Table 4). Subsequent screening of a cohort of 99 subjects (including healthy controls, MPN patients, and pediatric patients with essential thrombocythemia or hereditary erythrocytosis) detected JAK2 E846D substitution in 1 CALR⁺/V617F⁺ MPN patient and JAK2 R1063H in 1 V617F⁺ PV patient (supplemental Table 5) and in 1 triple-negative MPN patient (J.D.M.F. and R.K., unpublished data, August 29, 2015). The E846D substitution was previously reported in normal-karyotype acute myeloid leukemia¹³  and the JEG-3 cancer cell line¹⁴ ; R1063H was reported in 3 JAK2 V617F⁺ PV patients,²  as well as in acute myeloid leukemia¹³  and B-cell acute lymphoblastic leukemia cases.¹⁵  Altogether, these data suggest a much higher frequency of these 2 variants in MPN patients than in normal individuals.

Because the germ line JAK2 mutations emerged as the main disease-causing event in our patient, we created a cellular model: Ba/F3-EPOR cells expressing JAK2_E846D and JAK2_R1063H. Ba/F3-EPOR cells transfected with the wild-type human JAK2 (JAK2_wt) and V617F JAK2 were used as a negative and positive control, respectively.¹⁶  As expected, the JAK2_V617F transfectants exhibited survival advantage and supported constitutive proliferation in EPO-free or EPO-limiting conditions^1,4,17  (supplemental Figure 3A; Figure 2A). The viability of JAK2_E846D and JAK2_R1063H transfectants in EPO-free medium was comparable to the JAK2_wt transfectants (supplemental Figure 3A); neither these transfectants expressing individual JAK2 genetic variants nor double transfectants with both E846D and R1063H (not shown) became growth factor independent. Nevertheless, JAK2_E846D–expressing cells showed significantly increased proliferation in EPO-limiting conditions compared with JAK2_wt (Figure 2A). Improved proliferation for JAK2_R1063H transfectants was observed only in 0.1 U/mL EPO concentration (Figure 2A). These results suggested that E846D and R1063H substitutions have a weaker effect on the survival and proliferation capacity of cells than the oncogenic V617F mutation, which is consistent with their germ line transmission and with nonclonal hematopoiesis observed in the JAK2 E846D⁺ mother (described in supplemental Materials and methods). This also showed a stronger effect of E846D substitution on improved proliferation in EPO-limiting conditions compared with R1063H.

We then focused on JAK2 downstream signal transduction pathways. Upon interleukin-3 (IL-3) starvation and short incubation in IL-3–free EPO-containing medium, JAK2_E846D– and JAK2_R1063H–expressing cells showed increased STAT5 phosphorylation (Figure 2B). Weak activation of STAT1 was also observed for both mutants (supplemental Figure 3B).

Next, we tested the effect of our JAK2 mutants on STAT transcriptional activity by dual-reporter luciferase assay in JAK2-deficient γ-2A cells.¹⁸ Figure 2Ci shows that JAK2 E846D and R1063H significantly increased STAT5 transcriptional activity in EPO-free and low EPO conditions in the presence of EPOR when compared with cells expressing JAK2 wt. In EPO-limiting conditions, the coexpression of these 2 JAK2 mutants further improved STAT5 transcriptional activity compared with cells expressing a single JAK2 mutant (E846D or R1063H). No effect of JAK2 E846D and R1063H was observed on basal STAT5 activation via MPL or the granulocyte colony-stimulating factor (G-CSF) receptor (Figure 2Cii-iii) or on the activation of other STAT members (supplemental Figure 4). This was in agreement with increased STAT5 phosphorylation in the patient’s BM detected by immunohistochemical analysis of STAT5 (supplemental Figure 5) and is consistent with the erythroid phenotype in our patient.

To assess the effect of JAK2 mutants on primary cells, BM cells from C57BL/6 mice were retrovirally transduced with murine JAK2 expression vectors and used for in vitro colony assay. The cells expressing the E846D mutant formed significantly increased numbers of colony forming unit-erythroid (CFU-E) and BFU-E, when compared with cells with JAK2 wt in EPO-free or EPO-limiting conditions (Figure 2D). The observation that the R1063H mutant induced CFU-E rather than BFU-E colony growth (see also primary father’s progenitors in supplemental Figure 6) is reminiscent of the weak activation of EPOR by the anemic strain of the Friend virus complex gp55 envelope protein (gp55-A), which also specifically supported CFU-E formation.^19,20  For the EPO-free condition, the co-occurrence of the 2 mutations was slightly more effective in colony growth than the single mutants alone (Figure 2D). This assay indicated that JAK2 E846D and R1063H mutations facilitate hypersensitive response of erythroid progenitors from the patient or patient’s parents to EPO.

In silico modeling using the x-ray structures of the inactive and active JH1 domain of JAK2^21,22  suggested that the R1063H mutation would be presumed to facilitate the active conformation of JH1, whereas the JAK2 E846D mutant would show slower return to the ground state after cytokine stimulation (supplemental Figure 7). The effect of E846D substitution was confirmed by immunoblots analysis in which IL-3–starved JAK2_E846D transfectants showed prolonged activation of STAT5 and extracellular signal-regulated kinase 1/2 (ERK1/2) compared with JAK2_wt after high-dose EPO stimulation and subsequent EPO withdrawal (Figure 2E) resembling signaling downstream EPOR gain-of-function mutants.²³  The prolonged STAT5 and ERK1/2 activation downstream JAK2_E846D can be effectively reduced by JAK2 inhibitors, ruxolitinib or AZ-960, but not by wortmannin, a phosphoinositide-3-kinase inhibitor (supplemental Figure 8).

In summary, we establish that E846D and R1063H are weakly activating mutations and conclude that the cooperation of these activating JAK2 mutations causes a polyclonal hereditary erythrocytosis with megakaryocytic atypia, clinically resembling JAK2 exon 12–positive PV.

The authors thank Michael Schuster and Donat Alpar (the Biomedical Sequencing Facility at CeMM) as well as Martin Tichy and Anna Lapcikova (University Hospital and Faculty of Medicine and Dentistry in Olomouc) for their technical assistance, and Srdan Verstovsek (MD Anderson, Houston, TX) for critical reading of the paper and discussions.

This work was supported by Czech Science Foundation Project P301/12/1503 (V.D.). J.F.K. was supported by CZ.1.07/2.3.00/30.0041. Support to S.N.C. was from the Fonds de la Recherche Scientifique–Fonds National de la Recherche Scientifique, Belgium, the Salus Sanguinis Foundation, the Action de Recherche Concertée projects MEXP31C1 and ARC10/15-027 of the University catholique de Louvain, Brussels, the Fondation contre le Cancer, Brussels, the Pôle d'Attraction Interuniversitaire Programs BCHM61B5, and Belgian Medical Genetics Initiative (BeMG). E.L. was supported by a Fonds pour la formation à la Recherche dans l'Industrie et dans l'Agriculture PhD fellowship. The Sonderforschungsbereich grant from the Austrian Science Fund F4702-B20 is acknowledged for its generous support to R.K. J.D.M.F. acknowledges support from L’ORÉAL Österreich, For Women in Science Fellowship.

Contribution: K.K. and M.H. designed and performed research, analyzed data, and wrote the paper; C.P. designed and performed research, analyzed data, and reviewed the paper; J.F.K. performed research, analyzed data, and reviewed the paper; D.P. recruited the patient and contributed to the editing of the paper; E.L. performed research and contributed to the writing of the paper; B.K. performed research and analyzed data; J.D.M.F. designed and performed research and analyzed data; F.S. analyzed data; R.K. contributed to study design, analyzed data, reviewed the paper, and provided financial support; S.N.C. contributed to study design, analyzed data, wrote the paper, and provided financial support; and V.D. designed the study, analyzed data, wrote the paper, and provided financial support.

Conflict-of-interest disclosure: The authors declare no competing financial interests.

Correspondence: Vladimir Divoky, Department of Biology, Faculty of Medicine and Dentistry, Palacký University, Hnevotinska 3, Olomouc, 775 15, Czech Republic; e-mail: vladimir.divoky@upol.cz; or Stefan N. Constantinescu, Signal Transduction and Molecular Hematology Unit, Ludwig Institute for Cancer Research and de Duve Institute, Université catholique de Louvain, Avenue Hippocrate 74, UCL 75-4, Brussels B-1200, Belgium; e-mail: stefan.constantinescu@bru.licr.org.